论文部分内容阅读
Microstructures are important in determining the hulk properties of materials.X-ray micro-CT is a sam pie-non-destructive tool in characterizing the material microstructure.However, due to unavoidable variations in equipment and environment, the acquired CT images often have a degree of temporal and spatial variations.A method was proposed to compensate such variations in this paper.This method fits the trend graph of the synchrotron radiation light using a least squares polynomial curve fitting and uses the fitted curve to compensate the X-ray light intensity variations in the vertical direction for each image.These corrected image data were used as input for CT image reconstruction.These reconstructed CT images were then used to predict the three dimensional compositional distributions via a multi-energy least square segmentation technique.